天天看點

開源項目(庫)之libcurl學習(一)

最近抽了點時間,将之前開發中使用到的一些開源庫進行了下總結,主要是為了回顧一下自己所使用的一些庫基礎知識,并且加深了解,在這些庫中,首先一個庫就是libcurl,這個庫很強大,當時在做openstack swift API時使用到了,這個庫一個輕量級的HTTP程式設計庫,裡面封裝了一套基于HTTP的上層應用協定的資料包的基本操作,其支援FTP,FTPS,TFTP,HTTP,HTTPS,GOPHER,TELNET,DICT,FILE和LDAP,跨平台,支援Windows,Unix,Linux等,線程安全,支援Ipv6。并且易于使用。下面就從安裝開始,之前的開發是在Windows下開發的,現在的工作環境換為了Linux,但是不要緊,安裝過程差不多,步驟如下:

1)下載下傳libcurl檔案,下載下傳指令:sudo wget http://curl.haxx.se/download/curl-7.35.0.tar.gz

2)使用指令解壓檔案:tar -zxvf curl-7.35.0.tar.gz

3)./configure --prefix=/usr/local/libcurl

4)make

5)make install

經過上面幾個步驟,libcurl庫基本上就安裝好了,下面我們就稍微寫過小測試程式來測試測試,代碼如下:

#ifndef __HTTP_CURL__H
#define __HTTP_CURL__H

#include <boost/smart_ptr.hpp>
#include <boost/xpressive/xpressive_dynamic.hpp>
#include <boost/typeof/typeof.hpp>
#include <curl/curl.h>
#include <string>
#include <string.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
using namespace std;
using namespace boost;
using namespace boost::xpressive;
#define MAX_BUFFERSIZE 1024*10
class HttpCurl
{
    public:
        HttpCurl()
        {
            conn = NULL;
            memset(errBuffer,0,sizeof(errBuffer));
        }
        ~HttpCurl()
        {
            curl_easy_cleanup(conn);
        }
        bool HttpCurlInit(string& context)
        {
            CURLcode code;
            string error;

            code = curl_global_init(CURL_GLOBAL_DEFAULT);
            if(CURLE_OK != code)
            {
                printf("Failed to global init default\n");
                return false;
            }

            conn = curl_easy_init();
            if(NULL == conn)
            {
                printf("Failed to create CURL\n");
                return false;
            }

            code = curl_easy_setopt(conn,CURLOPT_ERRORBUFFER,error.c_str());
            if(CURLE_OK != code)
            {
                printf("Failed to set error buffer\n");
                return false;
            }

            code = curl_easy_setopt(conn,CURLOPT_WRITEFUNCTION,HttpCurl::write);
            if(CURLE_OK != code)
            {
                printf("Failed to set write\n");
                return false;
            }
            code = curl_easy_setopt(conn,CURLOPT_WRITEDATA,&context);
            if(CURLE_OK != code)
            {
                printf("Failed to set write data\n");
                return false;
            }
            return true;
        }
        bool setUrl(string& url)
        {
            CURLcode code;
            code = curl_easy_setopt(conn,CURLOPT_URL,url.c_str());
            if(CURLE_OK != code)
            {
                printf("Failed to set URL\n");
                return false;
            }
            return true;
        }

        bool getHttpResponse()
        {
            CURLcode code;
            std::string error;
            code = curl_easy_perform(conn);
            if(CURLE_OK != code)
            {
                printf("Failed to get [%s]",error.c_str());
                return false;
            }
            return true;
        }

        static long write(void* data,int size,int nmemb,std::string& context)
        {
            long sizes = size*nmemb;
            std::string temp((char*)data,sizes);
            context += temp;
            return sizes;
        }

        bool save(const string& context,std::string filename)
        {
            CURLcode code;
            int retcode = 0;
            code = curl_easy_getinfo(conn,CURLINFO_RESPONSE_CODE,&retcode);
            if((CURLE_OK == code)&& retcode ==200)
            {
                int length = strlen(context.c_str());
                FILE* file = fopen(filename.c_str(),"w+");
                fseek(file,0,SEEK_SET);
                fwrite(context.c_str(),1,length,file);
                fclose(file);
                return  true;
            }
            return false;
        }
    private:
        CURL* conn;
        char errBuffer[MAX_BUFFERSIZE];
};
           

測試程式代碼如下:

int main()
{
    string context;
    HttpCurl curl;
    curl.HttpCurlInit(context);
    curl.setUrl("www.renren.com");
    curl.getHttpResponse();
    curl.save(context,"text.txt");
    return 0;
}
           

通過簡單的測試用例,可以看出libcurl使用的一些基本流程,接下來我們再看看一個使用libcurl寫成的單線程爬蟲程式,這個程式實作思路很簡單,有機會将其改成多線程,代碼如下:

class Spider
{
    public:
        Spider(shared_ptr<HttpCurl>& cul):httpCurl(cul)
        {
            urlSet.clear();
            finishUrlSet.clear();
        }
        ~Spider(){}

        bool init(std::string& context)
        {
            return httpCurl->HttpCurlInit(context);
        }

        void parseUrl(const string& context)
        {
            const string tag = "href";
            const string tag2 = "\"";
            string::size_type tempBegin,tempEnd,iter;
            tempBegin = tempEnd = 0;
            iter= context.find(tag);
            while(iter != string::npos)
            {
                tempBegin = context.find(tag2,iter);
                if(tempBegin != string::npos)
                {
                    ++tempBegin;
                    tempEnd = context.find(tag2,tempBegin);
                }
                if(tempEnd != string::npos && tempEnd > tempBegin)
                {
                    string url;
                    url.assign(context,tempBegin,(tempEnd-tempBegin));
                    urlSet.insert(url);
                }
                iter = context.find(tag,tempEnd);
            }
        }
        void filterUrl()
        {
            string tag = "http";
            urlSet_Iter iter = urlSet.begin();
            for(;iter != urlSet.end();)
            {
                string::size_type index = (*iter).find(tag);
                if(index == string::npos)
                    urlSet.erase(iter);
                iter++;
            }
        }

        bool write(const string& context,const string& filename)
        {
            return httpCurl->save(context,filename);
        }

        void start(std::string url,std::string& context)
        {
            httpCurl->setUrl(url);
            httpCurl->getHttpResponse();
            parseUrl(context);
            filterUrl();
        }

        void displayUrl()
        {
            urlSet_Iter iter = urlSet.begin();
            for(; iter != urlSet.end();++iter)
            {
                cout<<*iter<<endl;
            }
        }
        void loop(const std::string& url,std::string& context)
        {
            start(url,context);
            for(urlSet_Iter iter = urlSet.begin();iter != urlSet.end();++iter)
            {
                if(finishUrlSet.find(*iter) != finishUrlSet.end())
                    continue;
                printf("%s\n",(*iter).c_str());
                char filename[64];
                memset(filename,0,sizeof(filename));
                sprintf(filename,"%d.html",fileIndex++);
                context.clear();
                start(*iter,context);
                write(context,filename);
                finishUrlSet.insert(*iter);
            }
        }
    private:
        shared_ptr<HttpCurl> httpCurl;
        std::set<string> urlSet;
        std::set<string> finishUrlSet;
        typedef std::set<string>::iterator urlSet_Iter;

};
           

測試程式:

#include "curlTest.h"

int main()
{
    string context;
    shared_ptr<HttpCurl> curl(new HttpCurl());
    Spider spider(curl);

    spider.init(context);
    spider.loop("www.renren.com",context);
    return 0;
}
           

測試結果:

http://a.xnimg.cn/favicon-rr.ico?ver=3

http://a.xnimg.cn/n/core/res/certificate.jpg

http://a.xnimg.cn/wap/apple_icon_.png

http://app.renren.com

http://app.renren.com/?origin=40206

http://app.renren.com/activity/specialpage?activity=girlgame&origin=40240

http://app.renren.com/activity/specialpage?activity=newgame&origin=40238

http://app.renren.com/activity/specialpage?activity=sanguo&origin=40239

http://app.renren.com/list?category=10&type=1&origin=40060&menu=1

http://app.renren.com/list?category=10&type=1&origin=40065&menu=1

http://app.renren.com/list?category=10&type=1&origin=40207&menu=1

http://app.renren.com/list?category=11&type=1

http://app.renren.com/list?category=11&type=1&origin=3142&added=0

http://app.renren.com/list?category=11&type=1&origin=40131

http://app.renren.com/list?category=11&type=1&origin=40132

http://app.renren.com/list?category=11&type=1&origin=40189

http://app.renren.com/list?category=11&type=1&origin=50298

http://app.renren.com/list?category=12&type=1&origin=3142&added=0

http://app.renren.com/list?category=12&type=1&origin=40131

http://app.renren.com/list?category=12&type=1&origin=40132

http://app.renren.com/list?category=13&type=1

http://app.renren.com/list?category=13&type=1&origin=3142&added=0

http://app.renren.com/list?category=13&type=1&origin=40131

http://app.renren.com/list?category=13&type=1&origin=40132

http://app.renren.com/list?category=13&type=1&origin=40188

http://app.renren.com/list?category=13&type=1&origin=40189

http://app.renren.com/list?category=13&type=1&origin=50298

http://app.renren.com/list?category=14&type=1

http://app.renren.com/list?category=14&type=1&origin=3113

http://app.renren.com/list?category=14&type=1&origin=3142&added=0

http://app.renren.com/list?category=14&type=1&origin=40188

http://app.renren.com/list?category=14&type=1&origin=50298

http://app.renren.com/list?category=15&type=1&origin=3142&added=0

http://app.renren.com/list?category=15&type=1&origin=40131

http://app.renren.com/list?category=17&type=1&origin=3142&added=0

http://app.renren.com/list?category=19&type=1

http://app.renren.com/list?category=19&type=1&origin=3113

http://app.renren.com/list?category=19&type=1&origin=3142&added=0

http://app.renren.com/list?category=19&type=1&origin=40132

http://app.renren.com/list?category=19&type=1&origin=40188

http://app.renren.com/list?category=19&type=1&origin=50298

http://app.renren.com/list?category=20&type=1&origin=40066&menu=2

http://app.renren.com/list?category=20&type=1&origin=40210&menu=2

^C

總結

      本篇博文主要是安裝了下libcurl,并且在此基礎上實作了一個簡單的爬蟲程式,代碼結構很簡單,至于libcurl的一些接口介紹什麼的,大家可以去其官網上看,在這裡就略過了,在寫這個爬蟲程式時,一開始一心想用boost裡的regex做正規表達式比對url,但是中間遇到了問題,是以臨時自己寫了個提取url的parseURL,另外這個爬蟲程式使用的是單線程的,有機會的話,我會将其改成支援多線程,其實很多東西在于自己去實踐,才能體會到其中的奧秘,好了,這篇博文到此就結束了,多謝。

如果需要,請注明轉載,多謝

繼續閱讀